LINEAR ALGEBRA - Solved Exercises 1

(Real and complex finite-dimensional vector spaces)

Nicolae Cotfas, version 09 Feb. 2021 (for future updates see https://unibuc.ro/user/nicolae.cotfas/ )
A PARTIAL EVALUATION 1 B PARTIAL EVALUATION 2
A1 Find the values of λ ∈ R for which B1 a) Find
B = {(1, −1), (2, λ)} is a basis of R2 , and the  an orthonormal basis of
K = x = (x1 , x2 , x3 , x4 ) ∈ R4 | x2 +x3 +x4 = 0 .
coordinates of v = (4, −3) with respect to B. b) Describe explicitly the orthogonal projector
A2 In R3consider the subspaces P : R4 → R4 corresponding to K ⊂ R4 .
V = (x, y, z) ∈ R3 | x − 2z = 0 and c) Find an  orthonormal
 basis of  
   x1 x2  T
3  x − y + z = 0, H= A= ∈ M2×2 (R) A = A .

W = (x, y, z) ∈ R  2x − y − z = 0 .

a) Compute: dim V, dim (V ∩ W),
dim W, dim (V + W).
b) Is it true that R3 = V ⊕ W ?
A3 Let A : R3 → R3 be the linear map defined by
its action on the elements of the canonical basis:
Ae1 = e1 − e2 ,
Ae2 = e1 + 2e2 + 3e3 ,
Ae3 = e1 − 2e2 − e3 .
a) Give the matrix of A in the canonical basis.
b) Compute dim (KerA) and dim (ImA).
c) Decide whether R3 = KerA ⊕ ImA or not.
A4 Consider the basis B0 = {(1, i), (1, −i)} of C2
and the linear operator
A : C2 → C2 , A(z, w) = (z − iw, iz + w).
a) Find the change of basis matrix from the
canonical basis B of C2 to B0 .
b) Find the matrix of A with respect to B0 .
c) Determine An , for n ∈ N.
Some definitions, notations and results
In a vector space V over K = R or K = C,
1 span {v1 , v2 , ..., vn } = {α1 v1 +...+αn vn | αk ∈ K}.
• V = span {v1 , v2 , ..., vn } and
B = {v1 , v2 , ..., vn }
if • α1 v1 +α2 v2 +...+αn vn = 0
is a basis of V
α1 = α2 = ... = αn = 0
2 Canonical basis of
x3 x4
B2 a) Find the eigenvalues and eigenspaces of
A : R2 → R2 , A(x1 , x2 ) = (πx1 +x2 , πx2 ),
B : C2 → C2 , B(z1 , z2 ) = (iz1 +z2 , z2 ).
b) Investigate the existence of the diagonal form.
B3 Let 
0 1 0

A = 1 0 1 .
0 1 0
a) Find an orthogonal matrix S such
that ST AS is a diagonal matrix.
b) Find an explicit expression of An for n ∈ N.
B4 a) Find the adjoint U † of the operator
U : C2 → C2 , U (z1 , z2 ) = √12 (iz1 +z2 , −z1 − iz2 ).
b) Prove that U is a unitary operator.
c) Diagonalize U .
B5 Let t ∈ (0, 1) and u1 , u2 ∈ R3 be two orthogonal
unit vectors, hu1 , u2 i = 0, ||u1 || = ||u2 || = 1.
a) Prove that the operator A : R3 → R3 ,
Ax = (1−t)hu1 , xiu1 +thu2 , xiu2 .
is a self-adjoint operator.
b) Find the eigenvalues of A.
c) Prove that hx, Axi ≥ 0, for any x ∈ R3 .
B6 a) Assume that A is a self-adjoint (Hermitian)
operator on a Hilbert space H. Prove that
⇓ Ker A ⊥ Im A and H = Ker A⊕Im A.
b) Solve the Schrödinger equation
i dψ
dt = Hψ
• K2 ≡ {x = (x 1 , x2 ) | xk ∈ K}
 is {e 1 = (1,
 0),
 e2 = (1,  0)}.
 describing the time evolution of a qubit
2 x 1
 1 0 with the Hamiltonian
• K ≡ x= x ∈ K is e1 = , e2 = .
 
x2  k

0 1 2 2 0 −i
H : C → C , H = −ωσ2 = −ω ,
i 0
3 Change of basis matrix in K3 from the old basis
wavefunction
B = {v1 , v2 , v3 }to the new basis B0 = {v10 , v20 , v30 } is
 
ψ (t)

α11 α12 α13 v10 = α11 v1 +α21 v2 +α31 v3 ψ(t) = ψ0 (t)|0i+ψ1 (t)|1i ≡ 0
ψ1 (t)
S = α21 α22 α23 where v20 = α12 v1 +α22 v2 +α32 v3 and initial state  
α31 α32 α33 v30 = α13 v1 +α23 v2 +α33 v3 . 1 1
ψ(0) = ψ0 (0)|0i+ψ1 (0)|1i ≡ 5
√ .
3 2
4 Matrix of A : K → K3 relative to B = {e1 , e2 , e3 } is
 Prove that ||ψ(t)|| = 1, for any t ∈ R.
a11 a12 a13 Ae1 = a11 e1 +a21 e2 +a31 e3
A = a21 a22 a23 where Ae2 = a12 e1 +a22 e2 +a32 e3 hw,xi
a31 a32 a33 Ae3 = a13 e1 +a23 e2 +a33 e3 8 Orthogonal projection of x on w 6= 0 is Pw x = hw,wi w.
Matrix of A : K3 → K3 relative to B0 is A0 = S−1 AS. 9 Gram-Schmidt orthogonalization method:
5 Sum of subspaces {v1 , v2 , ..., vn } linearly independent set of vectors
• W1 +W2 = {w1 +w2 | w1 ∈ W1 , w2 ∈ W2 }.
⇓
• Wdenoted
1 +W2 is a direct sum
by W1 ⊕W2 ⇔ W1 ∩W2 = {0}. • {w1 , w2 , ..., wn }, where
• dim(W1 +W2 ) = dim W1 +dim W2 −dim(W1 ∩W2 ) w1 = v1 ,
hw1 ,v2 i
w2 = v2 − hw 1 ,w1 i
w1 ,
6 Kernel of A : V → W Ker A = {x ∈ V | Ax = 0}. ..............................
Image of A : V → W Im A = {Ax | x ∈ V}. wn = vn − hw hw1 ,vn i
1 ,w1 i
w1 −· · · − hwhwn−1 ,vn i
n−1 ,wn−1 i
wn−1 ,
For A : V → W, dim V = dim Ker A+dim Im A is an orthogonal set of vectors such that
2
7 Standard • in R : h(x1 , x2 ), (y1 , y2 )i = x1 y1 +x2 y2 , • span {w1 , w2 , ..., wk } = span {v1 , v2 , ..., vk },
inner 2 for any k ∈ {1, 2, ..., n}.
product • in C : h(x1 , x2 ), (y1 , y2 )i = x̄1 y1 + x̄2 y2 .
 10 Basis B = {v1 , ..., vn } is orthonormal if hvj , vk i = δjk . SOLUTIONS for PARTIAL EVALUATION 2
2
11 Dirac’s “bra-ket” notation (in the  case
 of C ) B1 a) K = { (x1 , x2 , x3 , −x2 − x3 ) | x1 , x2 , x3 ∈ R}
  y1
x1 hx|yi = (x̄1 x̄2 ) = x̄1 y1 + x̄2 y2 , = {x1 (1, 0, 0, 0)+x2 (0, 1, 0, −1)+x3 (0, 0, 1, −1))| xk ∈ R}
|xi ≡ , y2
x2     = span {v1  = (1, 0, 0, 0), v2 = (0, 1, 0, −1), v3 = (0, 0, 1, −1))}
hx| ≡ (x̄1 x̄2 ) |xihy| =
x1
(ȳ1 ȳ2 ) =
x1 ȳ1 x1 ȳ2
. Gram- w1 = v1 = (1,hw0,,v0,i0),

x2 x2 ȳ1 x2 ȳ2 ⇒ w2 = v2 − hw11,w21 i w1 = (0, 1, 0, −1),
Schmidt
w = v − hw1 ,v3 i w − hw2 ,v3 i w = (0, − 1 , 1, − 1 )

12 ||u|| = 1 ⇒ Pu = |uihu|. 3 3 hw1 ,w1 i 1 hw2 ,w2 i 2 2 2
Orthogonal projector is an orthogonal basis of K ⇒
{u 1 , ..., u m } orthonormal
13 basis in a subspace K ⊂ H ⇒ on K is PP : H → H, {u1 = (1, 0, 0, 0), u2 = (0, √12 , 0, − √12 ), u3 = (0, − √16 , √26 , − √16 )}
P = |uk ihuk |.
is an orthonormal basis of K.  
3 0 0 0
    
Eigenvalues x1 a11 a12 x1
14 A = 2 −1 −1
of A : C2 → C2 , x2 a21
 a22 x2  0
b) P = |u1 ihu1 |+|u2 ihu2 |+|u3 ihu3 | = 31  .
 
a11 −λ a12  0 −1 2 −1
are the roots λ of the equation   = 0.  
a21 a22 −λ ( ! ) 0 −1 −1 2
0 √12
 
Eigenspace corresponding to λ is { x | Ax = λx }. 1 0 0 0
c) , √1 , is an orthonormal basis of H.
15 0 0 2
0 0 1
The adjoint †
the operator hx, Ayi =
is A† satisfying for any x, y ∈ H. hA x, yi,
of A : H → H B2 a, b) π 1   
π−λ 1


A= ,    = 0 ⇒ λ1 = λ2 = π.
16 With respect to an orthonormal basis, 0 π 0 π−λ 
† T
†
the matrix A of the adjoint of A is A = Ā . A(x 1 , x 2 ) = π(x 1 , x 2 ) ⇒ (x 1 , x 2 ) ∈ { α(1, 0) | α ∈ R }.
A is not diagonalizable (there does not exist a basis
17 A:H→H
is Hermitian ⇔ hx, Ayi = hAx, yi,
for any x, y ∈ H. ⇔ A = A
† containing only eigenvectors of A).
(self-adjoint)
   
i 1  i−λ 1 
18 A ∈ Mn×n (R) is a symmetric matrix if AT = A. B = ,  = 0 ⇒ λ1 = i, λ2 = 1.
0 1 0 1−λ
A ∈ Mn×n (R) is an orthogonal matrix if AT A = I. B(z1 , z2 ) = i(z1 , z2 ) ⇒ (z1 , z2 ) ∈ { α(1, 0) | α ∈ C }.
A ∈ Mn×n (C) is a unitary matrix if A† A = I. B(z1 , z2 ) = 1(z1 , z2 ) ⇒ (z1 , z2 ) ∈ { α(1, 1−i) | α ∈ C }.
Matrix of B relative to {(1, 0), (1, 1−i)} is diagonal.
 
SOLUTIONS for PARTIAL EVALUATION 1 B3 a) −λ 1 0  √
 1 −λ 1  = 0 ⇒ λ1 = 0, λ2,3 = ± 2.
A1 α1 (1, −1)+α2 (2, λ) = (0, 0)  0 1 −λ ! 
⇓ only for λ 6= −2. !
α1 = α2 = 0 1 √ 1
√
α α
Ax = 0 ⇒ x = √2 0 , Ax = ± 2x ⇒ x = 2 ± 2 .
(4,−3) = α1 (1, −1)+α2 (2, λ) ⇒ α1 = 6+4λ 2+λ , α2 = 2+λ .
1
−1 1
 √   
A2 a) V = {(2z, y, z) | y, z ∈ R } 2 √1 √ 1 0 √0 0
= {y(0, 1, 0)+z(2, 0, 1) | y, z ∈ R } S = 21  √0 2 − 2 satisfies ST AS = 0 2 √0
W = {(2z, 3z, z) | z ∈ R } = {z(2, 3, 1) | z ∈ R } = V ∩W. − 2 1 1 0 0− 2
{(0, 1, 0), (2, 0, 1)} basis in V ⇒ dim V = 2. and is orthogonal, that is, ST S = I.
{(2, 3, 1)} basis in W ⇒ dim W = dim (V ∩W) = 1. b)  n  
0 √0 0 0 √0 0
dim(V +W) = dim V +dim W −dim(V ∩W) = 2. An = S 0 2 √ 0  ST = S 0 ( 2)n √0  ST = ...
b) dim(V +W) = 2 ⇒ R3 6= V ⊕ W. 0  0 − 2 0  0 (− 2)n
A3 a) 
1 1 1

i 1 −i −1
Ae1 = e1 − e2 + 0 e3 , B4 U = √12 ⇒ U† = √12 ⇒ UU† = I.
Ae2 = e1 + 2e2 + 3e3 , ⇒ A = −1 2 −2 ⇒ −1 −i 1 i
Ae3 = e1 − 2e2 − e3 . 0 3 −1  c)
      
 √i −λ √1
 Matrix √
of U relative
√
to
x1 1 1 1 x1 x1 +x2 +x3

(1,( 2−1)i) (1,−( 2+1)i)
 2 2 =0 ⇒ λ
1,2 = ±i. √ √ , √ √
A x2  = −1 2 −2 x2  = −x1 +2x2 −2x3  ,  −12 −i
−λ
 √ √

4−2 2 4+2 2
2
x3 0 3 −1 x3 3x2 −x3 is a diagonal matrix.
A(x1 , x2 , x3 ) = (x1 +x2 +x3 , −x1 +2x2 −2x3 , 3x2 −x B5 a) We extend {u1 , u2 } up to an orthonormal basis
 3 ). 3
  x 1 +x 2 +x 3 = 0   B = {u1, u2 , u3 } of R . Matrix of A relative to B is
1−t 0 0

b) Ker A = (x1 , x2 , x3 )  −x1 +2x2 −2x3 = 0
  3x2 −x3 = 0
 A = 0 t 0= A† . b) Eigenvalues of A are 1−t, t and 0.
0 0 0
= {α(−4, 1, 3) | α ∈ R } ⇒ dim Ker A = 1. c) hx, Axi = (1−t)|hu1 , xi|2 +t|hu2 , xi|2 ≥ 0 for any x ∈ R3 .
dim Im A = dim R3 −dim Ker A = 3−1 = 2 ⇒ ∈ Ker A,
B6 x Ay ∈ Im A ⇒ hx, Ayi = hAx, yi = 0 ⇒ Ker A ⊥ Im A ⇒
Im A = span {Ae1 , Ae2 } = {α(1, −1, 0)+β(1, 2, 3) |α, β ∈ R}
c) (−4, 1, 3) 6∈ Im A ⇒ Ker A ∩ Im A = {(0, 0, 0)} Ker A∩Im A= {0}. But,
 dim Ker  A+dim  Im  A = dim H.
0 0

⇒ R3 = Ker A⊕ Im A.  i dψ ψ 0 0 −i ψ 0 ψ 0 ωψ1
=
dt = Hψ ⇒ i ψ 0 = −ω ⇒ 0
1 = −ωψ0

1 1 1 i 0 ψ 1 ψ
A4 Change of basis matrix from B  to B0 isS = .
i −i
 00
ψ0 +ω 2 ψ0 = 0
   
1 −i ψ0 (t) C1 sin ωt+C2 cos ωt
⇒ ⇒ =
b) Matrix of A relative to B is A =
i 1
. ψ00 = ωψ1 ψ1 (t) C1 cos ωt−C2 sin ωt !
√2 sin ωt+ √1 cos ωt
   
1 ψ (t)

2 0 ψ(0) = √15 ⇒ ψ(t) = 0 = 25 5
Matrix of A relative to B0 is A0 = S−1 AS = . 2 ψ1 (t) √ cos ωt− √1 sin ωt
.
c) 0 0 5 5
A2 (z, w) = A(z−iw, iz+w) = 2(z−iw, iz+w) = 2A(z, w),  2  2
2 3 2 n n−1 n−2 2 n−1 ||ψ(t)|| 2
= √2 sin ωt+ √1 cos ωt + √2 cos ωt− √1 sin ωt = 1.
⇒ A = 2A, A = 2 A, ... , A = AA = 2 A = 2 A. 5 5 5 5